Please note: The algorithm descriptions in English have been automatically translated. Errors may have been introduced in this process. For the original descriptions, go to the Dutch version of the Algorithm Register.
Prevention & enforcement: risk-oriented research
- Publication category
- Impactful algorithms
- Impact assessment
- DPIA
- Status
- In use
General information
Theme
Begin date
Contact information
Responsible use
Goal and impact
Enforcement includes all measures and activities the SVB takes to prevent errors and misuse by citizens and others. If this fails, we try to detect, correct and, if necessary, fine these errors. Prevention and customisation are paramount here. Risk-oriented research helps by identifying, investigating and analysing risks of illegality. The aim is to reduce these risks. Data analysis makes research more focused. Risk-based investigation is an important tool to determine in which situations there may be illegality. If a citizen falls into a risk profile, he or she is more likely to be selected for further investigation.
Considerations
This is an impactful algorithm that identifies risks with risk profiles and recommends measures to mitigate them. The law states what data we are allowed to use. A process description has been made for the programme and a DPIA is still to come. Each risk profile is checked by lawyers before use. There is also an ethical test that is completed separately for each risk profile.
Human intervention
If the results of the algorithm relate to a small group of citizens, a random sample is taken within that group. The selected files are further investigated by SVB staff. For each investigation, the employee decides which resources are appropriate to use. The investigations check whether the actual situation matches the data known to the SVB.
Risk management
We test our risk profiles in several ways. These include using the human rights assessment framework of the Human Rights Board to check whether the signal on which the profile is based contains discriminatory elements. In addition, each risk profile undergoes a legal test and an ethical test before use. Citizens covered by the risk profile are randomly selected for further investigation by supervisors. Based on the outcome of the examination, a decision is made by the service team regarding the impact on benefits.
Legal basis
Enforcement duties within the General Old Age Pensions Act (AOW), General Child Benefit Act (AKW), General Survivors Act (Anw), the General Income Support for the Elderly Act (AIO). Article 34 SUWI Act. Article 6(1)(e) AVG.
Impact assessment
Operations
Data
Risk profiles are almost always based on people's behaviour, such as removals, and not based on personal characteristics. No (directly) discriminatory selection criteria are used, such as racial or ethnic origin, nationality, political views, gender, sexual preference, religious or philosophical beliefs, criminal data and medical characteristics. The only exception is age, as there is an objective justification for this in some cases. Age is therefore not used in all signals, but an objective justification is considered for each signal. The risk profile is always reviewed by a lawyer before use. However, selection criteria that are (directly) discriminatory can be used for bias testing. The selection criteria are not the same for every signal and can be diverse. Some examples of selection criteria are:
- Law(s)
- Type of law
- Number of persons registered at an address
- Registered with third party/parties
- Presence/non-presence of a parent-child relationship at an address
- Having a correspondence address or letter address
- Being married/unmarried
- Having a partner abroad
- Having/not having income
- No persons paying a personal contribution in connection with receiving care from the Long-Term Care Act (WLZ). They stay in a care institution.
- No persons staying in a nursing home. Of these persons, it can be said with certainty where they reside.
- Person was not investigated by P&H last x to avoid selecting someone disproportionately often for investigation.
Technical design
Decision tree (rule-based).
Similar algorithm descriptions
- This algorithm helps Customs to select goods for inspection based on risk. It uses declaration data from companies and considers whether or not there are risks of inaccuracies in the declarations.Last change on 28th of February 2024, at 12:29 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- The risk prioritisation instrument (RPI) scores companies based on objectified compliance behaviour, environmental relevance and time lapse since last inspection. The higher a company is on the list, the higher the estimated probability of violations and the impact of any occurrences on people and the environment.Last change on 2nd of July 2024, at 10:17 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- The risk prioritisation scores companies based on compliance behaviour, environmental variables and the latest controls. The higher a company is on the list, the earlier it is eligible for inspections which can have an impact on any occurrences on humans and the environment.Last change on 3rd of April 2024, at 13:50 (CET) | Publication Standard 1.0
- Publication category
- Other algorithms
- Impact assessment
- Field not filled in.
- Status
- In use
- The algorithm is used to make an automated risk assessment for all Applications for Fixed Chargesgemoetkoming, prior to automated or manual granting and payment of the advance.Last change on 30th of May 2024, at 12:57 (CET) | Publication Standard 1.0
- Publication category
- High-Risk AI-system
- Impact assessment
- Field not filled in.
- Status
- In use
- The risk model (the algorithm) helps choose which NOW applications to investigate further. It gives clues as to whether the information given in the NOW application is correct. With these indications from the risk model, SZW reviews the application.Last change on 10th of June 2024, at 13:42 (CET) | Publication Standard 1.0
- Publication category
- Impactful algorithms
- Impact assessment
- Field not filled in.
- Status
- In use